A method for measuring speed by UWB multi-base station communication

Abstract

The application provides a UWB multi-base station communication speed measurement method, comprising the following steps: deploying a plurality of UWB base stations along an indoor track test line; installing a vehicle-mounted computer and a UWB tag, and integrating an IMU module; synchronizing the clocks of all UWB base stations and the vehicle-mounted computer, and performing bidirectional TOF ranging; pre-processing UWB signals and IMU data; selecting at least three base stations with the strongest signals, and performing fusion positioning based on a TDOA_AOA algorithm; and calculating instantaneous speed based on continuous positioning results, and performing dynamic error compensation. The application breaks through the precision bottleneck of dynamic speed measurement of long-coupling trains in a closed environment, and provides core data support for intelligent driving.

Description

Technical Field

[0001] This invention relates to the field of railway locomotive speed measurement technology, and in particular to a UWB multi-base station communication speed measurement method. Background Technology

[0002] With the intelligent development of rail transit, dynamic speed measurement technology on test and commissioning lines has become a core component in ensuring the safe operation of trains. In GPS-denied environments such as indoor test lines and tunnels, high-precision real-time speed measurement systems are required to control train start-stop, braking, and curve-passing performance. Traditional speed measurement solutions face challenges such as signal attenuation, multipath interference, and dynamic response delays in complex indoor environments, making it difficult to meet the centimeter-level speed measurement accuracy and millisecond-level real-time requirements of modern rail transit testing scenarios.

[0003] In existing technologies, such as the UWB-based intelligent rail vehicle depot positioning system and method provided by patent CN114449442A, multiple base stations are deployed in enclosed scenarios such as garages, and two-dimensional vehicle positioning is achieved through TOF ranging and central calculation. This solution has a certain accuracy in low-speed static scenarios, but it is difficult to achieve accurate positioning during high-speed speed measurement. Patent CN111970631A provides an underground train positioning and speed measurement system and method combining ultra-wideband and inertial navigation, but multipath interference from metal track reflections causes large fluctuations in UWB ranging. Furthermore, due to the fixed LMS step size factor, it cannot adapt to high-dynamic scenarios such as emergency braking of trains. Actual measurements show that convergence delays occur when acceleration changes abruptly. Additionally, UWB tags and INS devices are only deployed at the front of the train, resulting in velocity direction deviations between the front and rear when the train turns, without considering the motion state of the rear. Summary of the Invention

[0004] This invention addresses the aforementioned problems and aims to provide a UWB multi-base station communication speed measurement method. It reduces positioning errors through TDOA-AOA joint positioning, employs a third-order Lanczos differentiator for accurate instantaneous speed calculation, suppresses attitude errors in long train formations through head-to-tail speed fusion verification, and reduces metal reflection errors in UWB signals through SAGE algorithm multipath suppression. This method overcomes the bottleneck in dynamic speed measurement accuracy for long train formations in enclosed environments, providing core data support for intelligent driving.

[0005] Specifically, the first aspect of the present invention provides a UWB multi-base station communication speed measurement method, including the following steps:

[0006] Step 1: Deploy several UWB base stations equipped with directional antennas at preset intervals along the indoor track test line;

[0007] Step 2: Install onboard computers and UWB tags at the front and rear of the train, and integrate IMU modules;

[0008] Step 3: Synchronize the clocks of all UWB base stations and vehicle-mounted computers, and perform bidirectional TOF ranging between UWB tags and base stations;

[0009] Step 4: Preprocess the UWB signal and IMU data;

[0010] Step 5: Select at least three base stations with the strongest signals and perform fusion positioning based on the TDOA_AOA algorithm;

[0011] Step 6: Based on the continuous positioning results, the instantaneous velocity is calculated using a third-order Lanczos differentiator, and dynamic error compensation is performed.

[0012] Furthermore, the clock synchronization includes microsecond-level clock synchronization of all UWB base stations and vehicle-mounted computers via the PTP protocol.

[0013] A PTP master clock is deployed at the main base station, broadcasting synchronization messages via Ethernet. The onboard computer and the slave base station act as PTP slave clocks, recording message transmission and reception timestamps in hardware. Clock offset is compensated by calculating bidirectional transmission delay to eliminate clock drift errors between base stations and ensure TDOA ranging accuracy. Through hardware timestamps and the master-slave clock mechanism, synchronization accuracy can be controlled at the sub-microsecond level. In high-speed train scenarios, such as at 200 km / h, a 1-microsecond time error can lead to a 5.5-centimeter distance error. The calculation of clock offset in bidirectional transmission is as follows:

[0014] ;

[0015] in: This is due to clock skew;

[0016] Master clock message transmission time;

[0017] To receive time from the clock message;

[0018] The time is sent from the clock message;

[0019] This is the master clock message reception time.

[0020] Furthermore, step four includes processing the UWB channel impulse response using the SAGE algorithm and applying a fourth-order Butterworth low-pass filter to the IMU data.

[0021] The SAGE algorithm is used to process the UWB channel impulse response. First, the UWB channel impulse response CIR needs to be collected. After iteratively separating the multipath components, the direct wave is determined. If the time delay difference is greater than 10ns, the direct wave is retained; otherwise, the reflected wave is discarded. In the strong reflection environment of metal track, it can suppress the ranging fluctuation caused by multipath interference.

[0022] A fourth-order Butterworth low-pass filter is applied to the IMU data. The cutoff frequency threshold of the fourth-order Butterworth low-pass filter is set to 10Hz~30Hz. Since the main vibration frequency of the subway track is below 30Hz, the cutoff frequency can be flexibly adjusted according to the actual situation to effectively filter out track vibration noise.

[0023] Furthermore, the selection of the three base stations with the strongest signals must be based on base stations with RSSI > -80dBm and SNR > 15dB.

[0024] Selecting base stations with RSSI > -80dBm is intended to exclude weak signal base stations (in tunnel obstruction scenarios); selecting base stations with SNR > 15dB is intended to avoid high-noise base stations (in high-voltage electromagnetic interference areas). Base station data threshold selection: Actual test data from rail trains shows that when RSSI < -80dBm, the ranging error increases sharply from 10cm to over 50cm; when SNR < 15dB, the bit error rate exceeds... This will have a significant impact on the accuracy of distance measurement.

[0025] Furthermore, step five includes the following steps:

[0026] Step 5.1: Obtain the TOF ranging value and AOA angle measurement value of each selected base station, and obtain the location coordinates of the corresponding base station;

[0027] Obtain the TOF ranging values ​​of the three optimal base stations after filtering based on RSSI and SNR, and simultaneously obtain the AOA angle measurement values ​​of these base stations (provided by directional antennas). Retrieve the precise coordinate positions of the corresponding base stations from the system database (pre-calibrated during base station deployment).

[0028] Step 5.2: Construct a TDOA hyperbola using the base station with the strongest signal as a reference point to obtain the preliminary location area of ​​the target tag;

[0029] Using the base station with the strongest signal as a reference point, and utilizing the distance difference between other base stations and the reference base station (i.e., the TDOA measurement result), a cluster of hyperbolas is drawn on a two-dimensional plane: each hyperbola represents a set of points that meet a specific distance difference condition, and the intersection area of ​​multiple hyperbolas is the area where the target tag may exist. The intersection area of ​​multiple hyperbolas is used as the initial positioning area of ​​the target tag.

[0030] Step 5.3: Based on the preset direction vector of the indoor track test line, perform AOA angle constraints to compress the initial positioning area into a linear candidate area along the track;

[0031] Based on the preset direction vector of the track test line (such as 30° due north), the AOA measurement value of each base station is converted into a direction ray;

[0032] The specific physical constraints imposed on the orbit are: limiting the solution space to within ±2° of the orbital direction; and trunculating invalid solutions that exceed the orbital range (such as lateral offsets).

[0033] The output is compressed from an elliptical region into a linear candidate region along the orbit.

[0034] Step 5.4: Based on the comprehensive matching degree, find the optimal estimated position within the linear candidate region;

[0035] Within the linear candidate region of the track, dense sampling points (e.g., 10 points per meter) are set up to discretize the continuous track into a high-density grid, providing a basis for refined scoring, and a matching score is calculated for each sampling point:

[0036] ;

[0037] ;

[0038] ;

[0039] in: For distance matching degree;

[0040] This is the actual distance measured;

[0041] This represents the geometric distance from the point to the base station;

[0042] This represents the maximum tolerance deviation for the distance.

[0043] For angle matching degree;

[0044] This is the AOA measurement value;

[0045] This is the azimuth angle of the point relative to the base station;

[0046] This represents the maximum tolerance deviation for the angle.

[0047] This is a comprehensive score.

[0048] The point with the highest overall score is selected as the optimal estimated location.

[0049] Step 5.5: Project the obtained optimal estimated position coordinates vertically onto the center line of the track for correction to obtain the fused positioning coordinates.

[0050] Project the coordinates obtained in step 5.4 vertically onto the center line of the track, and calculate the offset distance between the projected point and the original point: if the offset is >0.1m, trigger a positioning anomaly alarm; if there is no anomaly, output the fused positioning coordinates on the track.

[0051] Furthermore, step six includes:

[0052] Step 6.1: Obtain at least three consecutive positioning coordinates of the vehicle, process them using a third-order Lanczos differentiator, and obtain the initial instantaneous velocity value at the current moment;

[0053] The positioning coordinates of the target vehicle at three consecutive time points (t-2, t-1, t) are obtained. These three position points are processed by a third-order Lanczos differentiator to output the initial instantaneous velocity value at the current moment (the initial velocity value at this time contains high-frequency noise).

[0054] Step 6.2: Input the instantaneous initial velocity value, IMU acceleration data, and historical velocity sequence into the adaptive Kalman filter, dynamically adjust the filter parameters, and obtain a smooth velocity estimate;

[0055] The instantaneous initial velocity value, IMU acceleration data, and historical velocity sequence are input into the adaptive Kalman filter. The filter parameters are dynamically adjusted according to the current acceleration amplitude. In the high acceleration state (acceleration and deceleration process), the process noise tolerance is increased, and in the steady state (uniform velocity process), the weight of the observed data is increased, and a smooth velocity estimate is output.

[0056] Step 6.3: Perform physical constraint verification on the smoothing speed estimate. If the verification passes, perform head and tail data verification; otherwise, trigger the three-level anomaly handling mechanism.

[0057] Step 6.4: Output the final velocity value and record the physical constraint verification results of this velocity calculation.

[0058] Furthermore, the dynamic adjustment of filter parameters includes: increasing the process noise tolerance when the absolute value of acceleration exceeds a preset range; and increasing the weight of the observation data when the absolute value of acceleration is within the preset range, i.e., in a stable state.

[0059] Furthermore, the physical constraint verification of the smoothed velocity estimate includes checking whether the angular deviation between the velocity direction and the track tangent and the rate of change of acceleration exceed the set values.

[0060] Check the angular deviation between the velocity direction and the track tangent. If the angular deviation is >2°, it is judged as an abnormal value. Detect the rate of change of acceleration. If the rate of change of acceleration is >0.3g / s, it is judged as jitter interference.

[0061] Furthermore, the head and tail data verification includes: synchronously acquiring the speed calculation results of the head and tail of the vehicle, performing a two-way data consistency check, and if the difference between the head and tail speed data is less than a threshold, then performing weighted fusion output; otherwise, initiating data comparison diagnosis and outputting high-confidence data.

[0062] The difference between the head and tail velocity data is less than the threshold, that is:

[0063] When driving in a straight line, the speed difference between the front and rear must be less than or equal to 0.1 m / s;

[0064] When driving on a curve, the speed difference between the front and rear must be less than or equal to 0.2 m / s;

[0065] For emergency braking, the speed difference between the front and rear ends must be less than or equal to 0.3 m / s;

[0066] If the conditions are met, a weighted fusion output is performed, with different weights assigned to different driving routes:

[0067] Straight-line driving: The weight of the front of the car is 0.8, and the weight of the rear of the car is 0.2;

[0068] Cornering: The weight of the front of the car is 0.75, and the weight of the rear of the car is 0.25;

[0069] Driving in tunnels: The weight of the front of the vehicle is 0.7, and the weight of the rear of the vehicle is 0.3;

[0070] Otherwise, initiate data comparison and diagnostics, and output high-confidence data:

[0071] The formula for comparing IMU confidence levels is as follows:

[0072] ;

[0073] in: Confidence level;

[0074] The function is for finding the minimum value;

[0075] The differential deviation of the IMU's acceleration and position;

[0076] The deviation between the angular velocity and trajectory curvature of the IMU;

[0077] If the confidence level of the front data is greater than that of the rear data + 0.2, output the front data; otherwise, output the rear data.

[0078] Furthermore, the three-level exception handling mechanism includes:

[0079] When the data changes instantaneously beyond the set threshold, the IMU-dominated mode is activated to perform first-level compensation, which is to perform velocity integral compensation based on the three-axis acceleration of the IMU.

[0080] When the sudden change in velocity / acceleration exceeds 3σ (σ is the standard deviation of the data in the first 10 seconds), first-level compensation is performed. The three-axis acceleration of the IMU is taken, the orbital tangential projection is performed on the three-axis acceleration, and velocity integral compensation is performed:

[0081] ;

[0082] in: Let be the velocity at time t;

[0083] The velocity at time t-1;

[0084] The projected acceleration of the IMU;

[0085] Let t be the time difference between time t and time t-1.

[0086] When the first-level compensation fails, the second-level compensation is performed, which is to extrapolate compensation based on the historical data sequence and by judging the current motion pattern.

[0087] When the first-level compensation exceeds the tolerance for two consecutive frames, and the compensation is insufficient to make the data jump back to the set threshold instantaneously, or when the IMU hardware fails and cannot perform speed compensation, the second-level compensation is performed. Based on the magnitude of acceleration of motion modes such as uniform speed, uniform acceleration, and uniform deceleration in the historical data sequence, and matching the historical data with the current motion mode, the current speed is extrapolated and compensated.

[0088] A fault alarm will be triggered when the secondary compensation fails or the data exceeds the safety boundary.

[0089] A fault alarm will be triggered when the extrapolated speed deviates from the track speed limit by more than 15% or when the acceleration direction is at an angle greater than 5° to the track (risk of derailment). Attached Figure Description

[0090] To more clearly illustrate the technical solutions in the embodiments of this drawing or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this drawing. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0091] Figure 1 This is a flowchart of the steps of the present invention.

[0092] The purpose, features, and advantages of this accompanying drawing will be further explained in conjunction with the embodiments and with reference to the accompanying drawing. Detailed Implementation

[0093] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments provided by this invention without inventive effort are within the scope of protection of this invention.

[0094] Obviously, the accompanying drawings described below are merely some examples or embodiments of the present invention. Those skilled in the art can apply the present invention to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this invention, modifications to design, manufacturing, or production based on the technical content disclosed in this invention are merely conventional technical means and should not be construed as insufficient disclosure of the present invention.

[0095] Unless otherwise specified, all embodiments and optional embodiments of the present invention can be combined with each other to form new technical solutions.

[0096] Unless otherwise specified, all technical features and optional technical features of this invention can be combined to form new technical solutions.

[0097] Unless otherwise specified, all steps of the present invention may be performed sequentially or randomly, preferably sequentially. For example, the method includes steps (a) and (b), indicating that the method may include steps (a) and (b) performed sequentially, or it may include steps (b) and (a) performed sequentially. For example, the mention that the method may also include step (c) indicates that step (c) may be added to the method in any order. For example, the method may include steps (a), (b), and (c), or it may include steps (a), (c), and (b), or it may include steps (c), (a), and (b), etc.

[0098] Unless otherwise specified, the terms "comprising" and "including" as used in this invention can be open-ended or closed-ended. For example, "comprising" and "including" can mean that other components not listed may also be included, or that only the listed components may be included.

[0099] Unless otherwise specified, the term "or" is inclusive in this invention. For example, the phrase "A or B" means "A, B, or both A and B". More specifically, the condition "A or B" is satisfied by any of the following conditions: A is true (or exists) and B is false (or does not exist); A is false (or does not exist) and B is true (or exists); or both A and B are true (or exist).

[0100] To better understand the solutions of the embodiments of the present invention, some related terms and concepts that may be involved in the embodiments of the present invention will be introduced below.

[0101] (1) UWB (Ultra-Wideband) is a carrier-free communication technology that uses non-sinusoidal narrow pulses in the nanosecond to picosecond range to transmit data. It features high-precision positioning, strong anti-interference capability, and strong multipath resolution.

[0102] (2) IMU (Inertial Measurement Unit) is a sensor composed of a three-axis accelerometer and a gyroscope, used to measure the angular velocity and linear acceleration of an object and to calculate its motion trajectory.

[0103] (3) Third-order Lanczos Differentiator: A numerical differentiation algorithm based on the Lanczos method, used for high-precision estimation of the time derivative of a signal (such as acceleration and jerk). In IMU data processing, it can improve the estimation accuracy of motion state (velocity and acceleration) and is suitable for high dynamic systems (such as UAV navigation) that require low latency and high frequency computing.

[0104] In this embodiment, as Figure 1 As shown, a UWB multi-base station communication speed measurement method includes the following steps:

[0105] Step 1: Deploy several UWB base stations equipped with directional antennas at preset intervals along the indoor track test line;

[0106] Step 2: Install onboard computers and UWB tags at the front and rear of the train, and integrate IMU modules;

[0107] Step 3: Synchronize the clocks of all UWB base stations and vehicle-mounted computers, and perform bidirectional TOF ranging between UWB tags and base stations;

[0108] Step 4: Preprocess the UWB signal and IMU data;

[0109] Step 5: Select at least three base stations with the strongest signals and perform fusion positioning based on the TDOA_AOA algorithm;

[0110] Step 6: Based on the continuous positioning results, the instantaneous velocity is calculated using a third-order Lanczos differentiator, and dynamic error compensation is performed.

[0111] Furthermore, clock synchronization includes microsecond-level clock synchronization of all UWB base stations and vehicle-mounted computers via the PTP protocol.

[0112] In this embodiment, a 1588v2 clock is used. A PTP master clock is deployed at the main base station, and synchronization messages are broadcast via Ethernet. The onboard computer and the slave base station act as PTP slave clocks. Hardware records the message transmission and reception timestamps. Clock offset is compensated by calculating bidirectional transmission delay to eliminate clock drift errors between base stations and ensure TDOA ranging accuracy. Through hardware timestamps and the master-slave clock mechanism, synchronization accuracy can be controlled at the sub-microsecond level. In high-speed train scenarios, such as at a speed of 200 km / h, a 1 microsecond time error will result in a 5.5 cm distance error. The calculation of clock offset in bidirectional transmission is as follows:

[0113] ;

[0114] in: This is due to clock skew;

[0115] Master clock message transmission time;

[0116] To receive time from the clock message;

[0117] The time is sent from the clock message;

[0118] This is the master clock message reception time.

[0119] Further, step four includes processing the UWB channel impulse response using the SAGE algorithm and applying a fourth-order Butterworth low-pass filter to the IMU data.

[0120] The SAGE algorithm is used to process the UWB channel impulse response. First, the UWB channel impulse response CIR needs to be collected. After iteratively separating the multipath components, the direct wave is determined. If the time delay difference is greater than 10ns, the direct wave is retained; otherwise, the reflected wave is discarded. In the strong reflection environment of metal track, it can suppress the ranging fluctuation caused by multipath interference.

[0121] A fourth-order Butterworth low-pass filter is applied to the IMU data. The cutoff frequency threshold of the fourth-order Butterworth low-pass filter is set to 10Hz~30Hz. Since the main vibration frequency of the subway track is below 30Hz, the cutoff frequency can be flexibly adjusted according to the actual situation to effectively filter out track vibration noise.

[0122] Furthermore, the three base stations with the strongest signals should be selected, specifically those with RSSI > -80dBm and SNR > 15dB.

[0123] Selecting base stations with RSSI > -80dBm is intended to exclude weak signal base stations (in tunnel obstruction scenarios); selecting base stations with SNR > 15dB is intended to avoid high-noise base stations (in high-voltage electromagnetic interference areas). Base station data threshold selection: Actual test data from rail trains shows that when RSSI < -80dBm, the ranging error increases sharply from 10cm to over 50cm; when SNR < 15dB, the bit error rate exceeds... This will have a significant impact on the accuracy of distance measurement.

[0124] Furthermore, step five includes the following steps:

[0125] Step 5.1: Obtain the TOF ranging value and AOA angle measurement value of each selected base station, and obtain the location coordinates of the corresponding base station;

[0126] Obtain the TOF ranging values ​​of the three optimal base stations after filtering based on RSSI and SNR, and simultaneously obtain the AOA angle measurement values ​​of these base stations (provided by directional antennas). Retrieve the precise coordinate positions of the corresponding base stations from the system database (pre-calibrated during base station deployment).

[0127] Step 5.2: Construct a TDOA hyperbola using the base station with the strongest signal as a reference point to obtain the preliminary location area of ​​the target tag;

[0128] Using the base station with the strongest signal as a reference point, and utilizing the distance difference between other base stations and the reference base station (i.e., the TDOA measurement result), a cluster of hyperbolas is drawn on a two-dimensional plane: each hyperbola represents a set of points that meet a specific distance difference condition, and the intersection area of ​​multiple hyperbolas is the area where the target tag may exist. The intersection area of ​​multiple hyperbolas is used as the initial positioning area of ​​the target tag.

[0129] Step 5.3: Based on the preset direction vector of the indoor track test line, perform AOA angle constraints to compress the initial positioning area into a linear candidate area along the track;

[0130] Based on the preset direction vector of the track test line (such as 30° due north), the AOA measurement value of each base station is converted into a direction ray;

[0131] The specific physical constraints imposed on the orbit are: limiting the solution space to within ±2° of the orbital direction; and trunculating invalid solutions that exceed the orbital range (such as lateral offsets).

[0132] The output is compressed from an elliptical region into a linear candidate region along the orbit.

[0133] Step 5.4: Based on the comprehensive matching degree, find the optimal estimated position within the linear candidate region;

[0134] Within the linear candidate region of the track, dense sampling points (e.g., 10 points per meter) are set up to discretize the continuous track into a high-density grid, providing a basis for refined scoring, and a matching score is calculated for each sampling point:

[0135] ;

[0136] ;

[0137] ;

[0138] in: For distance matching degree;

[0139] This is the actual distance measured;

[0140] This represents the geometric distance from the point to the base station;

[0141] This represents the maximum tolerance deviation for the distance.

[0142] For angle matching degree;

[0143] This is the AOA measurement value;

[0144] This is the azimuth angle of the point relative to the base station;

[0145] This represents the maximum tolerance deviation for the angle.

[0146] This is a comprehensive score.

[0147] The point with the highest overall score is selected as the optimal estimated location.

[0148] In this embodiment, the maximum tolerance deviation for distance is 1.5m, and the maximum tolerance deviation for angle is 10°.

[0149] Step 5.5: Project the obtained optimal estimated position coordinates vertically onto the center line of the track for correction to obtain the fused positioning coordinates.

[0150] Project the coordinates obtained in step 5.4 vertically onto the center line of the track, and calculate the offset distance between the projected point and the original point: if the offset is >0.1m, trigger a positioning anomaly alarm; if there is no anomaly, output the fused positioning coordinates on the track.

[0151] Furthermore, step six includes:

[0152] Step 6.1: Obtain at least three consecutive positioning coordinates of the vehicle, process them using a third-order Lanczos differentiator, and obtain the initial instantaneous velocity value at the current moment;

[0153] The positioning coordinates of the target vehicle at three consecutive time points (t-2, t-1, t) are obtained. These three position points are processed by a third-order Lanczos differentiator to output the initial instantaneous velocity value at the current moment (the initial velocity value at this time contains high-frequency noise).

[0154] Step 6.2: Input the instantaneous initial velocity value, IMU acceleration data, and historical velocity sequence into the adaptive Kalman filter, dynamically adjust the filter parameters, and obtain a smooth velocity estimate;

[0155] The instantaneous initial velocity value, IMU acceleration data, and historical velocity sequence are input into the adaptive Kalman filter. The filter parameters are dynamically adjusted according to the current acceleration amplitude. In the high acceleration state (acceleration and deceleration process), the process noise tolerance is increased, and in the steady state (uniform velocity process), the weight of the observed data is increased, and a smooth velocity estimate is output.

[0156] Step 6.3: Perform physical constraint verification on the smoothing speed estimate. If the verification passes, perform head and tail data verification; otherwise, trigger the three-level anomaly handling mechanism.

[0157] Step 6.4: Output the final velocity value and record the physical constraint verification results of this velocity calculation.

[0158] Furthermore, the filter parameters are dynamically adjusted, including: increasing the process noise tolerance when the absolute value of acceleration exceeds the preset range; and increasing the weight of the observation data when the absolute value of acceleration is within the preset range, i.e., in a stable state.

[0159] Furthermore, the smoothed velocity estimate is physically constrained and verified, including checking whether the angular deviation between the velocity direction and the track tangent and the rate of change of acceleration exceed the set values.

[0160] Check the angular deviation between the velocity direction and the track tangent. If the angular deviation is >2°, it is judged as an abnormal value. Detect the rate of change of acceleration. If the rate of change of acceleration is >0.3g / s, it is judged as jitter interference.

[0161] Furthermore, the head and tail data verification includes: synchronously acquiring the speed calculation results of the front and rear of the vehicle, performing a two-way data consistency check, and if the difference between the head and tail speed data is less than a threshold, then weighted fusion output is performed; otherwise, data comparison diagnosis is initiated, and high-confidence data is output.

[0162] The difference between the head and tail velocity data is less than the threshold, that is:

[0163] When driving in a straight line, the speed difference between the front and rear must be less than or equal to 0.1 m / s;

[0164] When driving on a curve, the speed difference between the front and rear must be less than or equal to 0.2 m / s;

[0165] For emergency braking, the speed difference between the front and rear ends must be less than or equal to 0.3 m / s;

[0166] If the conditions are met, a weighted fusion output is performed, with different weights assigned to different driving routes:

[0167] Straight-line driving: The weight of the front of the car is 0.8, and the weight of the rear of the car is 0.2;

[0168] Cornering: The weight of the front of the car is 0.75, and the weight of the rear of the car is 0.25;

[0169] Driving in tunnels: The weight of the front of the vehicle is 0.7, and the weight of the rear of the vehicle is 0.3;

[0170] Otherwise, initiate data comparison and diagnostics, and output high-confidence data:

[0171] The formula for comparing IMU confidence levels is as follows:

[0172] ;

[0173] in: Confidence level;

[0174] The function is for finding the minimum value;

[0175] The differential deviation of the IMU's acceleration and position;

[0176] The deviation between the angular velocity and trajectory curvature of the IMU;

[0177] If the confidence level of the front data is greater than that of the rear data + 0.2, output the front data; otherwise, output the rear data.

[0178] Furthermore, the three-tiered exception handling mechanism includes:

[0179] When the data changes instantaneously beyond the set threshold, the IMU-dominated mode is activated to perform first-level compensation, which is to perform velocity integral compensation based on the three-axis acceleration of the IMU.

[0180] When the sudden change in velocity / acceleration exceeds 3σ (σ is the standard deviation of the data in the first 10 seconds), first-level compensation is performed. The three-axis acceleration of the IMU is taken, the orbital tangential projection is performed on the three-axis acceleration, and velocity integral compensation is performed:

[0181] ;

[0182] in: Let be the velocity at time t;

[0183] The velocity at time t-1;

[0184] The projected acceleration of the IMU;

[0185] Let t be the time difference between time t and time t-1.

[0186] When the first-level compensation fails, the second-level compensation is performed, which is to extrapolate compensation based on the historical data sequence and by judging the current motion pattern.

[0187] When the first-level compensation exceeds the tolerance for two consecutive frames, and the compensation is insufficient to make the data jump back to the set threshold instantaneously, or when the IMU hardware fails and cannot perform speed compensation, the second-level compensation is performed. Based on the magnitude of acceleration of motion modes such as uniform speed, uniform acceleration, and uniform deceleration in the historical data sequence, and matching the historical data with the current motion mode, the current speed is extrapolated and compensated.

[0188] A fault alarm will be triggered when the secondary compensation fails or the data exceeds the safety boundary.

[0189] A fault alarm will be triggered when the extrapolated speed deviates from the track speed limit by more than 15% or when the acceleration direction is at an angle greater than 5° to the track (risk of derailment).

[0190] It should be noted that the present invention is not limited to the above-described embodiments. The above embodiments are merely examples, and any embodiments that have the same structure and perform the same effects as the technical concept within the scope of the present invention are included within the scope of the present invention. Furthermore, various modifications that can be conceived by those skilled in the art to the embodiments, and other ways of constructing by combining some of the constituent elements of the embodiments, without departing from the spirit of the present invention, are also included within the scope of the present invention.

Claims

1. A UWB multi-base station communication speed measurement method, characterized in that, The method comprises the following steps: Step 1: deploying a plurality of UWB base stations equipped with directional antennas at preset intervals along the indoor track test line; Step 2: installing a vehicle-mounted computer and a UWB tag at the head and tail of the train and integrating an IMU module; Step 3: synchronizing the clocks of all UWB base stations and vehicle-mounted computers, and performing bidirectional TOF ranging between the UWB tag and the base stations; Step 4: preprocessing the UWB signals and IMU data; Step 5: selecting at least three base stations with the strongest signals and performing fusion positioning based on a TDOA_AOA algorithm; The step 5 comprises the following steps: Step 5.1: obtaining the TOF ranging values and AOA angle measurement values of the selected base stations, and obtaining the position coordinates of the corresponding base stations; Step 5.2: taking the base station with the strongest signal as a reference datum to construct a TDOA hyperbola and obtain a preliminary positioning area of the target tag; Step 5.3: performing AOA angle constraint based on the preset direction vector of the indoor track test line to compress the preliminary positioning area into a linear candidate area along the track; Step 5.4: obtaining the optimal estimated position in the linear candidate area based on the comprehensive matching degree; Step 5.5: vertically projecting the obtained optimal estimated position coordinates onto the track center line for correction to obtain fusion positioning coordinates; Step 6: based on the continuous positioning results, using a third-order Lanczos differentiator to solve the instantaneous speed and performing dynamic error compensation.

2. The UWB multi-base station communication ranging method of claim 1, wherein, The clock synchronization comprises performing microsecond-level clock synchronization of all UWB base stations and vehicle-mounted computers through a PTP protocol.

3. The UWB multi-base station communication ranging method of claim 1, wherein, The step 4 comprises processing the UWB channel impulse response using a SAGE algorithm and applying a fourth-order Butterworth low-pass filter to the IMU data.

4. The UWB multi-base station communication ranging method of claim 1, wherein, The selection of the three base stations with the strongest signals requires selecting base stations with RSSI > -80 dBm and SNR > 15 dB.

5. The method of claim 1, wherein, The step 6 comprises: Step 6.1: obtaining at least three continuous positioning coordinates of the vehicle, processing the instantaneous speed initial value at the current time through a third-order Lanczos differentiator; Step 6.2: inputting the instantaneous speed initial value, IMU acceleration data, and historical speed sequence into an adaptive Kalman filter to dynamically adjust the filter parameters and obtain a smoothed speed estimate value; Step 6.3: performing physical constraint verification on the smoothed speed estimate value, and if the verification is passed, performing head-tail data verification, otherwise triggering a three-level abnormal processing mechanism; Step 6.4: outputting the final obtained speed value and recording the physical constraint verification situation of this speed solution.

6. The UWB multi-base station communication ranging method of claim 5, wherein, The dynamic adjustment of the filter parameters comprises: when the acceleration absolute value exceeds a preset interval, increasing the process noise tolerance; when the acceleration absolute value is within the preset interval, i.e. in a steady state, increasing the observation data weight.

7. The UWB multi-base station communication ranging method of claim 5, wherein, The physical constraint verification of the smoothed speed estimate value comprises: checking whether the angle deviation of the speed direction from the track tangent and the acceleration change rate exceed the set values.

8. The UWB multi-base station communication ranging method of claim 5, wherein, The head-tail data verification comprises: synchronously obtaining the speed solution results of the train head and tail, performing bidirectional data consistency verification, and if the head-tail speed data difference is less than a threshold, performing weighted fusion output; otherwise, starting data comparison diagnosis and outputting high-confidence data.

9. The method of claim 5, wherein, The three-level abnormality processing mechanism comprises: When the data instantaneous jump exceeds the set threshold, the IMU dominant mode is started, and the first-level compensation is performed, that is, the velocity integral compensation is performed based on the three-axis acceleration of the IMU; When the first-level compensation continuously fails, the second-level compensation is performed, that is, the extrapolation compensation is performed according to the historical data sequence by judging the current motion mode; When the second-level compensation fails or the data breaks through the safety boundary, the fault alarm is performed.

Images